On Sun, 2007-10-21 at 16:38 +0200, User 32X wrote:
Hello,
Francis Tyers wrote:
On Fri, 2007-10-19 at 18:23 +0200, User 32X
wrote:
The German Wikipedia is only about six times
larger
than the Volapük one. The later one is considerred
being small.
This is a ridiculous statement. If you take any metric apart from sheer
"page count" there is no comparison.
When I had a look at
http://www.wikipedia.org/
I had to asume that there was some kind of comparison.
There are better statistics here:
http://meta.wikimedia.org/wiki/List_of_Wikipedias
Articles Total Edits Admins Users Images
de 609770 1663501 35810679 283 425797 109499
vo 27686 30668 111338 5 173 368
But I saw some even better ones somewhere with statistics for article
depth (e.g. how many edits per article), etc.
Page/article count is only one of the things that needs to be taken into
account. For example, a wiki which has a page:edits ratio of 1:1 is not
likely to have a very developed community.
On the other
hand, between closely related languages, machine
translation is often very good, [...]
If there's actually a simple way of
machine translation
between two closely related languages, then why are
there more than one Wikipedias for that language family?
When a language community decides that it speaks a language, not a
dialect of another language, for whatever reason (for example national
identity) they may have a wish for a Wikipedia. Personally I don't think
this is up for you to decide.
Okay, there's the wish (and it's really not up for me
to decide), but why are they using the other content
then? Having the same content several times is just a
balkanization of the the community.
I agree :/
The Serbian Wikipedia works in Cyrillic and Latin
alphabets,
but there's still only one Wiki that serves both forms.
There are currently four Wikipedias for what until a few years ago was
called Serbo-Croatian. There is likely to be another one in the future.
I think it would be wonderful to see them merged. I'll let you get
started on that.
Nice try. See, the [[w:en:Little Rock Nine]] had to have
an armed escort to enter the Little Rock High School in
1957 – nearly a hundred years after the American Civil War.
The civil wars in the former Yugoslavia are only a decade
ago and there's still too much potential for hate between
these people. A combined Wikipedia would work for some of
them, but not for all. My hope is there'll be some day a
portal-like Wikipedia in serbo-croatia which takes the
content of the different Yugoslavian language versions and
combines them somehow. (With the ability to easily switch
between the same article in different languages.)
That would be a great idea, but as I said, I don't see it happening any
time soon :/
Maybe you
don't care about smaller languages, [...]
Believe me, I do. [...] These small
Wikipedias are often
interesting, but only when there's a comunity who writes
it. I don't need to read bot-created articles when there
are the same articles 1000 times better in other languages
I understand.
I don't see where I have suggested using a bot. Is that what you think I
have in mind? Machine translation needs to be post-editted, that is a
job for humans.
Okay, our positions move a bit closer together. After
the recent activities on vo.wikipedia I'm a bit cautious
when it comes to non-human content creation. As long as
there's a comunity to work on that content, this shouldn't
be that much of a problem. But still, if the text isn't
NPOV or based on OR this machine-based translations /could/
help to establish its contents.
Yes, quite right. A Wikipedia full of unchecked machine translations
does a lot more harm than good. The same goes for a Wikipedia full of
spam, or unchecked _human_ translations.
When a small Wikipedia doesn't have a special
field of
interest AND there is no large group of speakers, then it
is – in most cases – not worth supporting it.
What on earth do you mean by a "special field of interest" ?
The previously mentioned pdc.wikipedia (oviously) has a lot
of articles on German immigration in Pennsylvania. Many of
these articles aren't in the German nor the English Wikipedia.
These are the articles that make the pdc.wikipedia worth
reading, a dozen of these articles are more valuable than the
12,294 "articles" of German municipalities in the Volapük
Wikipedia.
I completely agree. Good articles are worth more than stubs, and that is
what machine translation is really about. Anyone can write a bot that
generates country/province stubs, even if they don't know the language.
But to translate anything more, for people who don't know the language,
machine translation can help. Providing:
1) The community assents, and there are guidelines for the process.
2) There are people there to post-edit the output.
You can read some discussion of the benefits and drawbacks here:
http://af.wikipedia.org/wiki/Wikipedia:Geselshoekie/MT
I'm glad that after a bit of a heated conversation we're largely in
agreement, and sorry if I came over a bit strong :)
Fran
Best regards,
32X